A double dose of climate research

Heidi Roop stands inside a share-edged snow trench holding a square-ended shovel aloft.
Heidi Roop

Predicting the course of climate change is tough. And without predictions, preparing for it isn’t easy, either.

To meet these challenges, the National Science Foundation (NSF) has invested $25 million in each of two new climate-oriented centers in which U of M researchers will play key roles. They are among six new NSF multi-institutional science and technology centers nationwide.

In the Center for Oldest Ice Exploration (COLDEX), researchers will drill miles down in Antarctica to study ice and trapped air to unearth an unprecedented record of links between greenhouse gases and climate in warmer zones and the long-term rhythms of Earth’s climate system. 

Heidi Roop, an assistant professor of climate science in the College of Food, Agricultural and Natural Resource Sciences and an Extension specialist, is COLDEX’s director for knowledge transfer.

In the center called Learning the Earth with Artificial Intelligence and Physics (LEAP), researchers will develop new algorithms to integrate observations from satellites and other sources to continually update models of how climate change affects different parts of the globe so people can prepare to weather the changes. Vipin Kumar, a Regents Professor in the College of Science and Engineering, leads the U of M team in this work. 

Deep data drilling

The oldest continuous record of Antarctic ice goes back about 800,000 years, but the team hopes to find a continuous record going back 1.5 million years, and possibly locate ice older than three million years.

“Our goal with COLDEX is to translate and share our discoveries so climate change knowledge belongs to everyone and can be used to take positive action across sectors,” Roop says.

The center’s work rests partly on the “rapid access ice drill,” co-developed by John Goodge, a professor at the U of M Duluth campus. It can cut through ice and rock to create an open borehole into which instruments can be lowered to measure properties and estimate age. 

“We completed successful field trials in early 2020 at a test site near McMurdo Station in Antarctica,” says Goodge. “It allowed us to recover ice and sub-glacial bedrock samples at the deepest levels yet achieved. The unique mobile drilling platform also proved to be more efficient and inexpensive than other types of drills.”

COLDEX is led by a team of researchers from Oregon State University.

Read more about COLDEX.

Soaring satellite science

Global climate models disagree on how much the planet will warm in the next 40 years, and on the severity of impacts like sea level rise and more frequent floods and droughts. Much of the problem comes down to trying to represent the details of complex physical and biological processes —like clouds reflecting sunlight into space, or trees and oceans absorbing carbon from the air—into the models. Processes interact, and many are poorly understood.

The U of M researchers will lead the development of a new generation of machine-learning algorithms that can detect, even with limited data, cause-and-effect relationships between various parts of the global climatic system. This will help scientists predict climate-driven changes in various parts of the planet.  

“These knowledge-guided machine learning techniques are fundamentally more powerful than standard black-box machine learning approaches that often fail to generalize to unseen scenarios,” Kumar says. “We believe that artificial intelligence combined with scientific knowledge has the potential to substantially improve climate models.”

LEAP will also train a new wave of students fluent in both climate science and working with big datasets and modern machine-learning algorithms.

LEAP is led by researchers from Columbia University.

Read more about LEAP.

Thu, 10/28/2021 - 14:42
A double dose of climate research
University of Minnesota, Twin Cities